The new 3DiVi Omni Platform 1.14.0 is packed with multiple enhancements and new features. Here's what you can expect:
Simplified Installation Process
The installation kit and algorithm have been updated. Now, you'll spend less time filling in configuration parameters for the distribution setup.
Self Managed Cluster (smc) Module
Installation and cluster configuration have been segregated into the smc module. This results in fewer dependencies and offers a clearer, more sequential path for administrators during setup.
Modular Product Deployment
Deployment has been divided into distinct modules - platform and image-api. Installation now occurs through a single entry point, with each step detailed clearly. These modular components will be easier to maintain in future releases, speeding up the delivery of new features to our users.
Enhanced Deployment Support
- Horizontal Pod Autoscaler: This allows automatic scaling of deployments, ensuring efficient cluster resource utilization.
- NodeSelector: Choose specific nodes for your deployments, optimizing resource allocation by placing high-demand services on fewer, more powerful nodes.
- Ingress Proxy Support: Image API services can now be utilized within the OMNI Platform distribution.
Amazon EKS Deployment
Deploy solutions on Amazon EKS, enabling automatic scaling of ImageAPI services in the Amazon cloud. This ensures efficient hardware resource utilization and reduces costs for solutions with fluctuating client/service loads.
AWS Module
Set up the environment for deploying the platform on AWS. Apart from auto-scaling, this module brings:
Simplified load-based scaling.
High database availability (backup, resilience, recovery capabilities).
Cloud Watch: Use AWS's built-in tools for logging and monitoring.
Flexible Configuration Profiles
Adjust Image API configuration object parameters, allowing the system to be tailored for specific use cases, such as safe city operations, access control systems, selfie identification, artistic photo processing, and more.
Enhanced Stability and Performance
After a series of quality and load tests, we've significantly improved platform and agent stability and resilience. Memory leaks have been addressed, and throughput has been increased for mass detection and profile creation.
External Database Support
Use an external database for more reliable data backup tools on a separate machine.
Updated Recognition Modules
The score calculation formula in the v12 recognition modules has been updated, offering a more user-friendly interpretation of face comparison results.